1、Designation: E2056 04 (Reapproved 2010)Standard Practice forQualifying Spectrometers and Spectrophotometers for Usein Multivariate Analyses, Calibrated Using SurrogateMixtures1This standard is issued under the fixed designation E2056; the number immediately following the designation indicates the ye
2、ar oforiginal adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon () indicates an editorial change since the last revision or reapproval.1. Scope1.1 This practice relates to the multivariate calibration
3、ofspectrometers and spectrophotometers used in determining thephysical and chemical characteristics of materials. A detaileddescription of general multivariate analysis is given in PracticeE1655. This standard refers only to those instances wheresurrogate mixtures can be used to establish a suitable
4、 calibra-tion matrix. This practice specifies calibration and qualificationdata set requirements for interlaboratory studies (ILSs), that is,round robins, of standard test methods employing surrogatecalibration techniques that do not conform exactly to PracticesE1655.NOTE 1For some multivariate spec
5、troscopic analyses, interferencesand matrix effects are sufficiently small that it is possible to calibrate usingmixtures that contain substantially fewer chemical components than thesamples that will ultimately be analyzed. While these surrogate methodsgenerally make use of the multivariate mathema
6、tics described in PracticesE1655, they do not conform to procedures described therein, specificallywith respect to the handling of outliers.1.2 This practice specifies how the ILS data is treated toestablish spectrometer/spectrophotometer performance qualifi-cation requirements to be incorporated in
7、to standard testmethods.NOTE 2Spectrometer/spectrophotometer qualification procedures areintended to allow the user to determine if the performance of a specificspectrometer/spectrophotometer is adequate to conduct the analysis so asto obtain results consistent with the published test method precisi
8、on.1.2.1 The spectroscopies used in the surrogate test methodswould include but not be limited to mid- and near-infrared,ultraviolet/visible, fluorescence and Raman spectroscopies.1.2.2 The surrogate calibrations covered in this practice are:multilinear regression (MLR), principal components regress
9、ion(PCR) or partial least squares (PLS) mathematics. Thesecalibration procedures are described in detail in PracticesE1655.1.3 For surrogate test methods, this practice recommendslimitations that should be placed on calibration options that areallowed in the test method. Specifically, this practice
10、recom-mends that the test method developer demonstrate that allcalibrations that are allowed in the test method producestatistically indistinguishable results.1.4 For surrogate test methods that reference spectrometer/spectrophotometer performance practices, such as PracticesE275, E925, E932, E958,
11、E1421, E1683,orE1944; TestMethods E387, E388,orE579; or Guide E1866, this practicerecommends that instrument performance data be collected aspart of the ILS to establish the relationship betweenspectrometer/spectrophotometer performance and test methodprecision.2. Referenced Documents2.1 ASTM Standa
12、rds:2D6277 Test Method for Determination of Benzene in Spark-Ignition Engine Fuels Using Mid Infrared SpectroscopyD6300 Practice for Determination of Precision and BiasData for Use in Test Methods for Petroleum Products andLubricantsE131 Terminology Relating to Molecular SpectroscopyE275 Practice fo
13、r Describing and Measuring Performanceof Ultraviolet and Visible SpectrophotometersE387 Test Method for Estimating Stray Radiant PowerRatio of Dispersive Spectrophotometers by the OpaqueFilter MethodE388 Test Method for Wavelength Accuracy of SpectralBandwidth of Fluorescence SpectrometersE579 Test
14、Method for Limit of Detection of Fluorescence ofQuinine Sulfate in SolutionE691 Practice for Conducting an Interlaboratory Study toDetermine the Precision of a Test MethodE925 Practice for Monitoring the Calibration of Ultraviolet-Visible Spectrophotometers whose Spectral Bandwidthdoes not Exceed 2
15、nm1This practice is under the jurisdiction of ASTM Committee E13 on MolecularSpectroscopy and Separation Science and is the direct responsibility of Subcom-mittee E13.11 on Multivariate Analysis.Current edition approved March 1, 2010. Published May 2010. Originallyapproved in 1999. Last previous edi
16、tion approved in 2004 as E2056 04. DOI:10.1520/E2056-04R10.2For referenced ASTM standards, visit the ASTM website, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For Annual Book of ASTMStandards volume information, refer to the standards Document Summary page onthe ASTM website.1C
17、opyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.E932 Practice for Describing and Measuring Performanceof Dispersive Infrared SpectrometersE958 Practice for Measuring Practical Spectral Bandwidthof Ultraviolet-Visible Spectrophotometers
18、E1421 Practice for Describing and Measuring Performanceof Fourier Transform Mid-Infrared (FT-MIR) Spectrom-eters: Level Zero and Level One TestsE1655 Practices for Infrared Multivariate QuantitativeAnalysisE1683 Practice for Testing the Performance of ScanningRaman SpectrometersE1866 Guide for Estab
19、lishing Spectrophotometer Perfor-mance TestsE1944 Practice for Describing and Measuring Performanceof Laboratory Fourier Transform Near-Infrared (FT-NIR)Spectrometers: Level Zero and Level One Tests3. Terminology3.1 Definitions:3.1.1 For definitions of terms and symbols relating toinfrared, ultravio
20、let/visible and Raman spectroscopy, refer toTerminology E131.3.1.2 For definitions of terms and symbols relating tomultivariate analysis, refer to Practices E1655.3.2 Definitions of Terms Specific to This Standard:3.2.1 spectrometer/spectrophotometer qualification, ntheprocedures by which a user dem
21、onstrates that the performanceof a specific spectrometer/spectrophotometer is adequate toconduct a multivariate analysis so as to obtain precisionconsistent with that specified in the test method.3.2.2 surrogate calibration, na multivariate calibrationthat is developed using a calibration set which
22、consists ofmixtures with pre-specified and reproducible compositions thatcontain substantially fewer chemical components than thesamples that will ultimately be analyzed.3.2.3 surrogate test method, na standard test method thatis based on a surrogate calibration.4. Summary of Practice4.1 Asurrogate
23、test method must specify the composition oftwo sets of samples. One set is used to calibrate thespectrometers/spectrophotometers. The second set of samplesis used to qualify the spectrometer/spectrophotometer to per-form the analysis. The compositions of both sets are expressedin terms of weight or
24、volume fraction depending on whetherthe samples are prepared gravimetrically or volumetrically. Thecompositions of both sets should be specified in the surrogatetest method. If the surrogate test method is being used toestimate a physical property, then the test method shouldindicate what value of t
25、he property is to be assigned to each ofthe calibration and qualification samples.4.2 The surrogate test method should specify the minimumspectrometer/spectrophotometer requirements for instrumentsthat can be used to perform the test method.4.3 The spectrometer/spectrophotometer test method shouldsp
26、ecify the exact conditions that are to be used to collect and,where appropriate, to calculate the spectral data used in thecalibration and analysis.4.4 The test method should specify the exact mathematicsthat are to be used to develop the multivariate calibration.Allowable spectral preprocessing met
27、hods should be defined.The specific mathematics (MLR, PCR or PLS) should bespecified, and the acceptable range for the numbers of variablesshould be given.4.5 When the ILS is conducted to establish the precision ofthe surrogate test method, the calibration data for all of theparticipating laboratori
28、es should be collected and used tocalculate a pooled standard error of calibration for the testmethod. The pooled standard error of calibration and itsassociated degrees of freedom should be reported in the testmethod.4.5.1 When a user is calibrating a spectrometer/spectrophotometer, the standard er
29、ror of calibration is calcu-lated and compared to the pooled standard error of calibrationfrom the ILS to determine if the performance of the calibratedspectrometer/spectrophotometer is adequate to produce analy-ses of the precision specified in the test method.4.5.2 If a user is purchasing a precal
30、ibrated spectrometer/spectrophotometer, the instrument vendor should supply thestandard error of calibration and its statistical comparison tothe pooled standard error of calibration.4.6 During the ILS, each participating laboratory analyzes aset of qualification samples and reports both the composi
31、tionsof the qualification set and the estimates made using themultivariate analysis. A pooled error of qualification is calcu-lated and reported as part of the test method along with itscorresponding degrees of freedom.4.6.1 Before a user may use the spectrometer/spectrophotometer, it must be qualif
32、ied to perform the surro-gate test method. The qualification set is analyzed, and astandard error of qualification is calculated. The standard errorof qualification is statistically compared with the pooledstandard error of qualification to determine if the performanceof the calibrated spectrometer/
33、spectrophotometer is adequate toproduce analyses of the precision specified in the test method.4.6.2 Spectrometer/spectrophotometer qualification is re-quired regardless of whether the calibration is performed by thevendor or the user.4.6.3 Spectrometer/spectrophotometer qualification shouldbe repea
34、ted after major maintenance has been performed on thespectrometer/spectrophotometer so as to determine whetherrecalibration is required.5. Significance and Use5.1 This practice should be used by the developer ofstandard test methods that employ surrogate calibrations.5.1.1 This practice assists the
35、test method developer insetting and documenting requirements for the spectrometer/spectrophotometers that can perform the test method.5.1.2 This practice assists the test method developer insetting and documenting spectral data collection and compu-tation parameters for the test method.5.1.3 This pr
36、actice assists the test method developer inselecting among possible multivariate analysis procedures thatcould be used to establish the surrogate calibration. Thepractice describes statistical tests that should be performed toensure that all multivariate analysis procedures that are allowedE2056 04
37、(2010)2within the scope of the test method produce statisticallyindistinguishable results.5.1.4 This practice describes statistical calculations that thetest method developer should perform on the calibration andqualification data that should be collected as part of the ILSthat establishes the test
38、method precision. These calculationsestablish the level of performance that spectrometers/spectrophotometers must meet in order to perform the testmethod.5.2 This practice describes how the person who calibrates aspectrometer/spectrophotometer can test the performance ofsaid spectrometer/spectrophot
39、ometer to determine if the per-formance is adequate to conduct the test method.5.3 This practice describes how the user of a spectrometer/spectrophotometer can qualify the spectrometer/spectrophotometer to conduct the test method.6. Surrogate Calibrations6.1 Practices E1655 assumes that the calibrat
40、ion set used todevelop a multivariate model contains samples of the sametype as those that are to eventually be analyzed using themodel. Practices E1655 requires use of outlier statistics toensure that samples being analyzed are sufficiently similar tothe calibration samples to produce meaningful re
41、sults. Forsome spectroscopic analyses, however, it is possible to cali-brate using gravimetrically or volumetrically prepared mix-tures that contain significantly fewer components than thesamples that will ultimately be analyzed. For these surrogatetest methods, the outlier statistics described in P
42、ractices E1655are not appropriate since all samples are expected to be outliersrelative to the simplified calibrations. Thus, surrogate testmethods cannot fulfill the requirements of Practices E1655.While surrogate test methods may make use of the mathemat-ics described in Practices E1655, they shou
43、ld not claim tofollow the procedures described in that practice.6.1.1 In developing surrogate test methods, it is necessary tothoroughly understand and account for potential spectral inter-ferences. Typically, the spectral range used in surrogate cali-brations will be limited so as to minimize inter
44、ferences. Forthose interferences that cannot be eliminated through limitingthe spectral range, representative components that mimic theinterference should be included in the calibration mixtures.6.1.2 Test Method D6277 provides an example of a surro-gate test method. The FT-MIR analysis of benzene i
45、n gasolineis calibrated using mixtures of benzene, isooctane, toluene andxylenes and PLS mathematics. The calibration mixtures con-tain far fewer components than gasoline, but the spectral rangeused in the analysis is limited to a narrow range about arelatively interference-free benzene peak. Toluen
46、e and xylenesare used in the calibration mixtures to adequately mimic theinterferences that are present in gasolines.6.2 Calibration Sets:6.2.1 The sets of surrogate samples that are used to calibratethe spectrometers/spectrophotometers should satisfy the re-quirements of Practices E1655.Ifk is the
47、number of variables(MLR wavelengths or frequencies, PCR principal componentsor PLS latent variables) used in the model, then the minimumnumber of calibration samples should be the greater of 24 or6k. If the calibration set is derived from an experimentaldesign, and if the spectra have been shown to
48、be linearfunctions of the component concentrations, then fewer calibra-tion samples can be used, but in all cases the minimum numberof calibration samples should be the greater of 24 or 4k. Theexperimental design must independently vary all componentsover the desired analysis range.6.2.2 When calibr
49、ating for a single component, the calibra-tion set should uniformly span the range over which theanalysis of that component is to be conducted. Additionalcomponents that are present in the calibration set to simulateinterferences should be independently and uniformly variedover a range at least as large as is likely to be encounteredduring actual application of the test method.6.2.3 When calibrating for a property that depends on morethan one chemical component, the calibration set shoulduniformly span the range over which the property analysis is tobe conducted, and all compon
